An intensive and systematic three-year follow-up evaluation will be conducted on a sample of severe conduct problem children and their families participating in a treatment study now in progress, that compares the immediate treatment effects, short term persistence effects, and a set of therapeutic process variables for parent training (n = 50) versus community treatments (n = 20). The present study will assess the long term effectiveness of the two treatment approaches in reducing the vulnerability of these high risk youngsters to subsequent patterns of conduct disorders, delinquency, psychiatric problems, substance (ab)use, academic failure, and poor social adjustment. In addition, four areas to parent family management skills (discipline, supervision, positive parenting, and problem solving) and parent self-confidence will be examined as long term treatment outcomes. Finally, four parent risk variables (psychiatric, criminal, substance, abuse, and stress) and their relation both to family management and to child risk variables will be studied. In addition to the two treated groups, an increased risk comparison (n = 26) and a none increased risk comparison (n = 26) will be included so that comparisons with high risk youngsters similar to the treated participants, but lacking systematic treatment, and low risk youngsters who are not expected to be particularly vulnerable or poorly adjusted, can be made. Data on parenting skills, parent risk variables, and child risk and adjustment variables will be collected from all four groups in each of the three follow-up years, thus producing a 4 x 3 design. Multivariate analysis of variance will be the primary form of statistical analysis in evaluating the long term outcome hypotheses. Multiple regression analysis will be used to evaluate the efficacy of the parenting skills and parent risk variables as predictors of unique variance in the child outcome variables. Finally, structural equation modeling will be used to test the adequacy of several theoretical models.